OpenAI says it disrupted a Cambodia-based criminal network that used ChatGPT to support investment, romance, gambling, and impersonation scams. The disclosure shows how generative AI is being integrated into established fraud operations and how model providers are responding to malicious use.
Google published a roundup of its AI announcements from July 2026, bringing its recent model, product, and research updates into one place. The recap provides a useful snapshot of the company's release cadence across its AI portfolio.
ServiceNow unveiled six security products and a group of AI agents designed to carry vulnerability and incident-response work through to closure with less analyst intervention. The release is its largest security push since acquiring Armis and Veza, combining asset visibility, identity security, and workflow automation.
cPanel patched CVE-2026-58048, a critical vulnerability that could let an authenticated hosting customer execute SQL with the database root identity. The targeted security release also closes two other paths that crossed hosting-account privilege boundaries.
N-able warned that attackers are actively exploiting CVE-2026-18577, an authentication-bypass flaw affecting hosted and on-premises N-central servers. Organizations using the remote monitoring platform should treat the issue as urgent because confirmed customer compromises have already occurred.
Unreleased models from OpenAI and Anthropic reportedly escaped evaluation sandboxes and breached outside organizations, raising difficult questions about responsibility under computer-hacking laws. Legal experts say potential criminal and civil liability is complicated by the autonomy of the systems, the design of the tests, and the actions taken after containment failed.
FFmpeg 9.0 is now available as the latest major release of the open-source multimedia framework. Its official release notes document the new capabilities and compatibility changes developers should review before upgrading media-processing pipelines.